In recent years, a communication network has a significant role as a social infrastructure that provides various services, and failure of the communication network has an incalculable impact on users. Therefore, health-checking of the communication network has become a very important issue.
Patent Literature 1 (International Publication WO2005/048540) discloses a technique that uses a keep-alive frame to detect a failure in a communication network. More specifically, in a communication system in which a plurality of base nodes perform communication through one or more relay node, each base node transmits a keep-alive frame that is broadcasted by the relay node. Here, the plurality of base nodes mutually transmit and receive the keep-alive frame and detect failure by monitoring arrival state of the keep-alive frame transmitted from the other side node. In this case, in order to health-check all physical links in the communication network, it is necessary to configure a plurality of communication routes so as to cover all the physical links and to transmit and receive the keep-alive frame with respect to each communication route. That is, it is required to transmit and receive a large number of keep-alive frames. This causes increase in transmission and reception burden placed on each base node.
Non-Patent Literature 1 (S. Shah and M. Yip, “Extreme Networks' Ethernet Automatic Protection Switching (EAPS) Version 1”, The Internet Society, October 2003; (http://tools.ietf.org/html/rfc3619).) discloses a health-check technique in a communication network that is configured in a ring shape. In this case, a plurality of switches are connected through communication lines to form a ring shape, and one health-check frame is transferred sequentially along the ring. For example, a master switch on the ring transmits the health-check frame from a first port. Another switch forwards the received health-check frame to the next switch. The master switch receives the self-transmitted health-check frame at a second port, and thereby can confirm that no failure occurs. This technique assumes such a ring-shaped network structure and thus is not versatile.
Patent Literature 2 (Japanese Patent No. 3740982) discloses a technique that a management host computer performs health-check of a plurality of host computers. First, the management host computer determines an order of the health-check for the plurality of host computers. Next, the management host computer generates a health-check packet into which a health-check table is incorporated. The health-check table has a plurality of entries respectively related to the plurality of host computers, and the plurality of entries are arranged in the above determined order. Each entry includes an address of the related host computer and a check flag. Then, the management host computer transmits the health-check packet to a first host computer. A host computer that receives the health-check packet searches for the related entry in the health-check table and marks the check flag of the corresponding entry. After that, the host computer refers to the address in the next entry and transmits the health-check packet to the next host computer. Due to repetition of the above-mentioned processing, one health-check packet travels the host computers. Eventually, the management host computer receives the health-check packet that has traveled in this manner. Then, the management host computer determines that a failure occurs in a host computer the corresponding check flag of which is not marked.
According to Patent Literature 3 (Japanese Patent Publication JP-2006-332787), one health-check packet travels a plurality of monitor-target terminals, as in the case of Patent Literature 2. A similar health-check table is incorporated into the health-check packet. However, each entry includes, instead of the above-mentioned check flag, a check list in which such information as a date and time and an operating status is to be written. A monitoring terminal transmits the health-check packet to a first monitor-target terminal. When receiving the health-check packet, the monitor-target terminal judges whether or not itself is operating normally. In a case of a normal operation, the monitor-target terminal searches for the related entry in the health-check table and writes designated information such as the date and time and the operating status in the check list of the corresponding entry. Then, the monitor-target terminal refers to the address in the next entry and transmits the health-check packet to the next monitor-target terminal. Here, if communication with the next monitor-target terminal is impossible, the monitor-target terminal transmits the health-check packet to the monitor-target terminal after the next monitor-target terminal. Due to repetition of the above-mentioned processing, one health-check packet travels the monitor-target terminals. Eventually, the monitoring terminal receives the health-check packet that has traveled in this manner. If the designated information is not written in any check list, the monitoring terminal determines that a failure occurs.
It should be noted that Patent Literature 4 (Japanese Patent Publication JP-2000-48003), Patent Literature 5 (Japanese Patent Publication JP-H8-286920), Patent Literature 6 (Japanese Patent Publication JP-H11-212959) and Patent Literature 7 (Japanese Patent Publication JP-H3-191464) describe a method for solving a traveling salesman problem.